Legged robots have the potential to traverse diverse and rugged terrain. To find a safe and efficient navigation path and to carefully select individual footholds, it is useful to be able to predict properties of the terrain ahead of the robot. In this letter, we propose a method to collect data from robot-terrain interaction and associate it to images. Using sparse data acquired in teleoperation experiments with a quadrupedal robot, we train a neural network to generate a dense prediction of the terrain properties in front of the robot. To generate training data, we project the foothold positions from the robot trajectory into on-board camera images. We then attach labels to these footholds by identifying the dominant features of the force–torque signal measured with sensorized feet. We show that data collected in this fashion can be used to train a convolutional network for terrain property prediction as well as weakly supervised semantic segmentation. Finally, we show that the predicted terrain properties can be used for autonomous navigation of the ANYmal quadruped robot.

Where Should I Walk? Predicting Terrain Properties From Images Via Self-Supervised Learning

With the popularity of using deep learning-based models in various categorization problems and their proven robustness compared to conventional methods, a growing number of researchers have exploited such methods in environment sound classification tasks in recent years. However, the performances of existing models use auditory features like log-mel spectrogram (LM) and mel frequency cepstral coefficient (MFCC), or raw waveform to train deep neural networks for environment sound classification (ESC) are unsatisfactory. In this paper, we first propose two combined features to give a more comprehensive representation of environment sounds Then, a fourfour-layer convolutional neural network (CNN) is presented to improve the performance of ESC with the proposed aggregated features. Finally, the CNN trained with different features are fused using the Dempster–Shafer evidence theory to compose TSCNN-DS model. The experiment results indicate that our combined features with the four-layer CNN are appropriate for environment sound taxonomic problems and dramatically outperform other conventional methods. The proposed TSCNN-DS model achieves a classification accuracy of 97.2%, which is the highest taxonomic accuracy on UrbanSound8K datasets compared to existing models.

https://www.mdpi.com/1424-8220/19/7/1733/pdf

Environment Sound Classification Using a Two-Stream CNN Based on Decision-Level Fusion.

Cyber-physical systems (CPS) are interconnected architectures that employ analog and digital components as well as communication and computational resources for their operation and interaction with the physical environment. CPS constitute the backbone of enterprise (e.g., smart cities), industrial (e.g., smart manufacturing), and critical infrastructure (e.g., energy systems). Thus, their vital importance, interoperability, and plurality of computing devices make them prominent targets for malicious attacks aiming to disrupt their operations. Attacks targeting cyber-physical energy systems (CPES), given their mission-critical nature within the power grid infrastructure, can lead to disastrous consequences. The security of CPES can be enhanced by leveraging testbed capabilities in order to replicate and understand power systems operating conditions, discover vulnerabilities, develop security countermeasures, and evaluate grid operation under fault-induced or maliciously constructed scenarios. Adequately modeling and reproducing the behavior of CPS could be a challenging task. In this paper, we provide a comprehensive overview of the CPS security landscape with an emphasis on CPES. Specifically, we demonstrate a threat modeling methodology to accurately represent the CPS elements, their interdependencies, as well as the possible attack entry points and system vulnerabilities. Leveraging the threat model formulation, we present a CPS framework designed to delineate the hardware, software, and modeling resources required to simulate the CPS and construct high-fidelity models that can be used to evaluate the system’s performance under adverse scenarios. The system performance is assessed using scenario-specific metrics, while risk assessment enables the system vulnerability prioritization factoring the impact on the system operation. The overarching framework for modeling, simulating, assessing, and mitigating attacks in a CPS is illustrated using four representative attack scenarios targeting CPES. The key objective of this paper is to demonstrate a step-by-step process that can be used to enact in-depth cybersecurity analyses, thus leading to more resilient and secure CPS.

/pdf/cyber-physical-energy-systems-security-threat-modeling-risk-i4m2rebetp.pdf

Cyber-Physical Energy Systems Security: Threat Modeling, Risk Assessment, Resources, Metrics, and Case Studies

This paper presents the design and implementation of a bounding controller for the MIT Cheetah 2 and its experimental results. The paper introduces the architecture of the controller along with the...

High-speed bounding with the MIT Cheetah 2: Control design and experiments

Power quality control based on voltage sag/swell, unbalancing, frequency, THD and power factor using artificial neural network in PV integrated AC microgrid

This paper proposes a novel control solution designed to solve the local and grid-connected distributed energy resources (DERs) management problem by developing a generalizable framework capable of controlling DERs based on forecasted values and real-time energy prices. The proposed model uses sampling-based model predictive control (SBMPC), together with the real-time price of energy and forecasts of PV and load power, to allocate the dispatch of the available distributed energy resources (DERs) while minimizing the overall cost. The strategy developed aims to find the ideal combination of solar, grid, and energy storage (ES) power with the objective of minimizing the total cost of energy of the entire system. Both offline and controller hardware-in-the-loop (CHIL) results are presented for a 7-day test case scenario and compared with two manual base test cases and four baseline optimization algorithms (Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Quadratic Programming interior-point method (QP-IP), and Sequential Quadratic Programming (SQP)) designed to solve the optimization problem considering the current status of the system and also its future states. The proposed model uses a 24-hour prediction horizon with a 15-minute control horizon. The results demonstrate substantial cost and execution time savings when compared to the other baseline control algorithms.

/pdf/sampling-based-model-predictive-control-of-pv-integrated-2f1l45kfli.pdf

Sampling-Based Model Predictive Control of PV-Integrated Energy Storage System Considering Power Generation Forecast and Real-Time Price

Energy-efficient navigation is an important technology for mobile robots because of its potential to increase the operation time of the robot. In particular, when coupled with a dynamic legged quadruped, the need for energy savings is made more apparent as payloads are limited. Due to the complexity in modeling motion and power models of these robots, a new approach is necessary to effectively motion plan for these complex robots. We accomplish this by using Sampling-Based Model Predictive optimization (SBMPO) which was extended for use on the LLAMA quadrupedal platform in simulation. SBMPO allows for direct generation of trajectories while using a heuristic-based search to speed up computations. This approach is shown to effectively motion plan while optimizing for energy consumption and maintaining the natural dynamics of the robot in a simulated environment.

Energy Efficient Navigation for Running Legged Robots

A structure engulfed in flames can pose an extreme danger for fire-fighting personnel as well as any people trapped inside. A companion robot to assist the fire-fighters could potentially help speed up the search for humans while reducing risk for the fire-fighters. However, robots operating in these environments need to be able to operate in very low visibility conditions because of the heavy smoke, debris and unstructured terrain. This paper develops an audio classification algorithm to identify sounds relevant to fire-fighting such as people in distress (baby cries, screams, coughs), structural failure (wood snapping, glass breaking), fire, fire trucks, and crowds. The outputs of the classifier are then used as alerts for the fire-fighter or to modify the configuration of a robot capable of navigating unstructured terrain. The approach used extracts an array of features from audio recordings and employs a single hidden layer, feed forward neural network for classification. The simplicity in network structure enables performance on limited hardware and obtains classification results with an overall accuracy of 85.7%.

Sound Identification for Fire-Fighting Mobile Robots

Motion planning for legged machines such as RHex-type robots is far less developed than motion planning for wheeled vehicles. One of the main reasons for this is the lack of kinematic and dynamic models for such platforms. Physics based models are difficult to develop for legged robots due to the difficulty of modeling the robot-terrain interaction and their overall complexity. This paper presents a data driven approach in developing a kinematic model for the X-RHex Lite (XRL) platform. The methodology utilizes a feed-forward neural network to relate gait parameters to vehicle velocities.

Kinematic modeling of a RHex-type robot using a neural network

Outdoor mobile robots currently treat vegetation as obstacles that need to be avoided. In order to have less conservative robots that fully exploit their motion capabilities, it is required to obtain models of the interaction of vegetation with the vehicle. This work proposes and experimentally verifies models of interaction of wheeled and tracked vehicles with pliable vegetation. In addition, a methodology to map perceptual features of the environment to resistive forces experienced by the robots is presented.

Mario Harper

Papers

Sampling-Based Model Predictive Control of PV-Integrated Energy Storage System Considering Power Generation Forecast and Real-Time Price

Energy Efficient Navigation for Running Legged Robots

Sound Identification for Fire-Fighting Mobile Robots

Kinematic modeling of a RHex-type robot using a neural network

Characterization and Traversal of Pliable Vegetation for Robot Navigation